Development of an Indoor Positioning & Fall Detection System Based on Wearable Devices
Date Issued
2016
Date
2016
Author(s)
Huang, Bo-Chen
Abstract
The elderly population in many developed countries are steadily growing, Taiwan has also entered the aging society. The change of population structure has made us gradually conscious of the aging problems. Therefore, there is no doubt that we need to pay more attention to the home-care for the elderly. As far as home-care issue is concerned, falling is one of the major accidents at home for the elder people. It not only causes lots of physiological and psychological injuries, but it is also the main obstacle for elder people to live alone. In this thesis, we provide an ICT solution to home-care of elderly people. The user will wear a smart watch which can detect the wrist motion instantaneously. The sensor data are delivered to the server, which is able to detect the motion in real-time. If a falling accident is detected, the system will send a short message to the caregiver. Then indoor positioning function provides the location of the user to the caregiver so that they can assist the elder user immediately. In order to validate our system, we implemented two functions separately in different platforms. In the part of fall detection, we took advantage of Android wear smart watch to collect motion data, which will be sent to the Django web server. Then, the laptop could compute and determine the detection result by means of MATLAB simulation. In the part of indoor positioning, we built a real-time system. Besides using the National Instrument software-defined radio platform to receive the RF signal, we also implemented hardware acceleration with FPGA. In this way, we can track motion of the elderly people on the monitor screen in real time.
Subjects
Home-care
Fall detection
Smart watch
Android wear
Indoor Positioning
Soft Defined Radio
Real-time System
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-105-R02943052-1.pdf
Size
23.32 KB
Format
Adobe PDF
Checksum
(MD5):ba1a5ddadf393fdf9d9c1904b13cb8b5